INTRODUCTION

Until recently, one of the more severe constraints that have hindered the implementation of MCG in practical clinical settings has been the need for measurements to be made within a magnetically shielded room [1]. However, recent advances in SQUID system technology such as improved noise suppression techniques, better field sensitivity and highly balanced gradiometer systems let us construct a SQUID device, which allows us to perform MCG measurements in a totally unshielded environment. We demonstrated this at several US hospitals. The goal was to verify the reliability and reproducibility of MCG measurements in typical hospital settings by collecting data from both healthy volunteers and patients with documented heart disease. More than 200 volunteers have been evaluated. We present and compare examples of MCG data from CAD patients and healthy subjects. The results indicate that an unshielded MCG system is a potentially powerful tool to detect electric abnormalities in the human heart. CardioMag Imaging (CMI, www.cardiomag.com) has manufactured five magnetocardiographs, which are currently ready for use for clinical pilot studies.

METHODS

The SQUID sensors are coupled to second-order wire-wound hardware gradiometers with a baseline of 5.5 cm and a magnetic field resolution of < 20 fT/√Hz (at signal frequencies above 10 Hz) without any magnetic room shielding. A novel proprietary method of protection against radio-frequency electromagnetic interference should permit one to use the system in many clinical locations. System operation is computer-controlled and largely automated. The signals are acquired and processed by unique software capable of signal filtering, signal averaging, heart current reconstruction, and derivation of diagnostic parameters. The instrument shown in Fig. 1 has 9 magnetic field sensors near the bottom of the housing close to the patient's torso. MCG data are acquired at 36 locations above the torso by making four sequential measurements in mutually adjacent positions as shown in Fig. 2. After obvious signal artifacts are removed, the data are filtered and averaged for all positions. The software then interpolates the information and calculates the complete cardiac magnetic field maps (CMFM) as shown in Fig. 3. The first row of Fig. 3 shows different stages of the cardiac cycle with the effective current dipole orientation indicated by an arrow (P: peak of atrial depolarization activity; Q: initial septal activation; R: peak of ventricular depolarization; T: peak of ventricular repolarization).

Figure 1. The CMI Magnetocardiograph installed in a hospital, without magnetic room shielding.

Figure 2. Relative positions of the heart (thick outline) and the nine sensors (small circles) inside the cryostat housing (large circle) at four consecutive positions over the torso.

Healthy subject

Patient 1, LCX (100% stenosis)

Patient 2, LAD (90% stenosis)

Patient 3, RCA (80% stenosis)

Patient 4, LAD (90%), RCA (90%), LCX (50%)

Figure 3. Cardiac Magnetic Field Maps of stages in the cardiac cycle for patients with varying degrees of stenosis. Arrows indicate the estimated direction of the current flow.

The second row shows the corresponding CMFM obtained from a healthy subject. The magnetic isolines are uniformly distributed indicating that the current is flowing undisturbed. All other rows display the CMFM's of patients with CAD.

As an example of MCG data we show in Fig. 3 CMFMs of several patients with documented stenosis in coronary arteries: LAD (left anterior descending), LCX (left circumflex), RCA (right coronary) and combinations thereof, and compare them to the healthy subject.

RESULTS

While all four patients had a normal 12-lead ECG, the CMI Magnetocardiograph data show abnormal patterns.

Patient 1 has an LCX stenosis. Atrial depolarization (P) and the start of septal activity (Q) appear usual. In both cases the orientation of the current is as in the healthy case and the field distribution is homogenous. At R both are changed: the current flows around a block inferred in the upper right part of the map. This might correspond to an area of low conductivity. It is located laterally in a region where the circumflex is normally located and supplies blood flow to the lateral wall of the left ventricle. T shows abnormal repolarization. This is indicated by the turn of the current flow by 90 degrees and the shift towards the left area of the map.

Patient 2 has an LAD stenosis. P looks almost normal although a slightly unusual field distribution at the lower right part of the map can be noticed. Q and R show ideal patterns whereas T is abnormal. An additional split in the negative pole appears in the lower right corner of the map, which is an indicator for a distorted repolarization process. The current flow is redirected probably due to an apparent conduction block, which is regionalized on the map to be near the left ventricular apex. The apex area is often perfused by the LAD artery.

Patient 3 has an RCA stenosis. P shows an unusual pattern by not being dipolar. The current orientation during septal activation is towards the right side of the heart, which suggests a conduction abnormality during early depolariztion. R, reflecting mostly left ventricular depolarization, looks normal. Interestingly, the T pattern is similar to that in patient 2, although the underlying disease is different.

Patient 4 has a three-vessel disease. P shows reversed depolarization indicating that the atrial pacing trigger might be firing and initiating depolarization from an abnormal location. In Q, the direction of the current appears normal but the flow is fragmented in a central portion of the map. During R, the current is oriented away from the left ventricular apex.

CONCLUSION

MCG data have been recorded with the CMI magnetocardiograph in hospital environments without any magnetic shielding. The data show that it is possible to identify differences between healthy and pathologic hearts (even when the 12-lead ECG is normal) by simply viewing CMFM at a few select points during the cardiac cycle. We plan to gather sufficient statistical information to correlate CMFMs with specific pathologies. In summary, we have demonstrated that we can obtain interesting information about the electrophysiology of the heart that may have clinical utility even in an unshielded room environment.

REFERENCES

[1] H. Koch, "SQUID MCG: Status and Perspectives", IEEE Trans. Appl. Supercond., vol. 11(1), pp. 49-59, 2001.

ACKNOWLEDGMENT

We are grateful to S. Repnoy of CMI and Y. Polyakov of RED, Inc. for untiring efforts in system design and construction. This research was supported in part by the Department of Energy, Basic Energy Sciences.